ResNet

PyTorch implementation of ResNet¹ as defined in Torchvision.

Pre-trained models

mozuma.models.resnet.pretrained.torch_resnet_imagenet

TorchResNetModule model pre-trained on ImageNet

Parameters:

Name	Type	Description	Default
resnet_arch	ResNetArchs	Identifier for the ResNet architecture to load. Must be one of: resnet18 resnet34 resnet50 resnet101 resnet152 resnext50_32x4d resnext101_32x8d wide_resnet50_2 wide_resnet101_2	required
device	torch.device	Torch device to initialise the model weights	device(type='cpu')
training_mode	TorchResNetTrainingMode | None	Whether to return features or labels in the forward function. Used for training when computing the loss.	None

Returns:

Type	Description
TorchResNetModule	A PyTorch ResNet module pre-trained on ImageNet

Base model

The ResNet model is an implementation of a TorchModel.

mozuma.models.resnet.modules.TorchResNetModule

PyTorch ResNet architecture.

See PyTorch's documentation.

Attributes:

Name	Type	Description
resnet_arch	ResNetArchs	Identifier for the ResNet architecture to load. Must be one of: resnet18 resnet34 resnet50 resnet101 resnet152 resnext50_32x4d resnext101_32x8d wide_resnet50_2 wide_resnet101_2
label_set	LabelSet	The output labels. Defaults to ImageNet 1000 labels.
device	torch.device	Torch device to initialise the model weights
training_mode	TorchResNetTrainingMode | None	Whether to return features or labels in the forward function. Used for training when computing the loss.

Pre-trained state origins

See the stores documentation for usage.

mozuma.models.resnet.stores.ResNetTorchVisionStore

Model store to load ResNet weights pretrained on ImageNet from TorchVision

---

1. Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), volume, 770–778. 2016. doi:10.1109/CVPR.2016.90. ↩